Weekly Report
Robot: EURJPY V1.27
Base version: EURJPY V1 branch
Test week: Week 27
Start date: 2026-06-29
End of review: 2026-07-04
Symbol: EURJPY
Timeframe: M5
Mode: PAPER_SIM_ONLY
Orders sent: NO
Test type: Live paper observation
Weekly research target: +5R net per week
Purpose of This Report
This report reviews the Week 27 live paper result of EURJPY V1.27 Soft Parallel 30/70.
EURJPY is treated as a separate research branch. It is not reviewed as a direct comparison against EURGBP or any other symbol. The purpose of this report is to decide whether the EURJPY V1 branch should continue, pause, or produce a next controlled test version.
The main question for Week 27 was:
Did EURJPY V1.27 produce enough positive R potential, and can the negative side be reduced with one clear filter?
Week 27 Final Result
EURJPY V1.27 closed 54 paper trades during the reviewed Week 27 data set.
Final Week 27 result:
Trades: 54
Plus trades: 24
Minus trades: 29
Zero trades: 1
Gross +R: +28.6968R
Gross -R: -29.0381R
Net R: -0.3412R
Average R per trade: -0.0063R
Winrate: 44.44%
TP: 16
SL: 26
MAX_HOLD: 12
The raw version finished close to breakeven, but it did not reach the weekly research target of +5R net.
Target Review
Weekly target: +5R net
Actual result: -0.3412R net
Target status: FAIL
EURJPY V1.27 was not accepted as a ready demo-stage candidate after Week 27 because the raw net result was below the weekly target.
However, the version was not rejected as useless.
The version produced a very large gross positive result:
Gross +R: +28.6968R
This means the EURJPY V1.27 logic found many profitable movement opportunities during the week.
The problem was that the negative side was almost the same size:
Gross -R: -29.0381R
The result was therefore not a lack of positive potential. The result was a balance problem between strong winners and too many losing trades.
Why EURJPY V1.27 Was Not Accepted as a Demo Candidate
EURJPY V1.27 was not moved to demo-stage because it did not reach the +5R weekly net target.
The raw result was:
Net R: -0.3412R
The version also had a high number of SL exits:
SL: 26
This made the version too unstable in its raw form.
A robot candidate with this structure cannot be treated as ready. It still needs a filter that can reduce negative exposure while keeping enough of the strong gross positive potential.
Why EURJPY V1.27 Was Not Rejected
EURJPY V1.27 was not rejected because its gross positive result was strong.
The version produced:
Gross +R: +28.6968R
This is a large positive R base for one week of live paper observation.
A weak logic usually fails because it cannot generate enough positive R. EURJPY V1.27 did generate enough positive R, but it also allowed too many bad entries.
This makes EURJPY V1.27 a research candidate for filtering, not a finished robot and not a dead branch.
Plus and Minus Trade Audit
After the Week 27 result, a plus/minus audit was performed.
The purpose of the audit was to identify where losing trades were concentrated.
The audit checked the original trade fields and additional M1, M5, and M15 entry context.
The goal was not to add many filters. The goal was to find one simple technical condition that separated a better group from a worse group.
The main audit question was:
Are the losing trades concentrated in a repeated entry condition?
Main Finding: M15 Directional Half
The strongest practical finding was related to the M15 candle context at entry time.
The audit found that many losing trades appeared when the M15 close was not in the directional half for the candidate trade direction.
This condition is represented as:
m15_close_in_directional_half = NO
The good group was:
m15_close_in_directional_half = YES
This means the M15 close was positioned in the half of the candle that supported the candidate direction.
For BUY candidates, the M15 close should be in the upper half.
For SELL candidates, the M15 close should be in the lower half.
Good Group: M15 Close in Directional Half
When the M15 close was in the directional half, the result was strong.
m15_close_in_directional_half = YES
Trades: 27
Plus trades: 17
Minus trades: 10
Gross +R: +22.8183R
Gross -R: -9.4532R
Net R: +13.3651R
Winrate: 62.96%
TP: 13
SL: 8
MAX_HOLD: 6
This group alone exceeded the weekly research target.
Weekly target: +5R
Filtered group result: +13.3651R
Target status: PASS
This was the strongest filter candidate found in the EURJPY V1.27 audit.
Bad Group: M15 Close Not in Directional Half
When the M15 close was not in the directional half, the result was strongly negative.
m15_close_in_directional_half = NO
Trades: 27
Plus trades: 7
Minus trades: 19
Zero trades: 1
Gross +R: +5.8785R
Gross -R: -19.5849R
Net R: -13.7064R
Winrate: 25.93%
TP: 3
SL: 18
MAX_HOLD: 6
This group contained a large part of the negative exposure.
The blocked group had many more SL exits than TP exits:
TP: 3
SL: 18
MAX_HOLD: 6
This confirms that the M15 directional half condition was not neutral. It separated a strongly negative group from a much better group during Week 27.
Historical Filter Test on Week 27
A historical Week 27 filter test was performed using the EURJPY V1.27 closed trade set.
The test rule was:
Allow only trades where m15_close_in_directional_half = YES.
Block trades where m15_close_in_directional_half = NO.
The result changed as follows.
Base EURJPY V1.27:
Trades: 54
Plus trades: 24
Minus trades: 29
Zero trades: 1
Gross +R: +28.6968R
Gross -R: -29.0381R
Net R: -0.3412R
Average R per trade: -0.0063R
Winrate: 44.44%
TP: 16
SL: 26
MAX_HOLD: 12
Filtered EURJPY V1.27:
Trades: 27
Plus trades: 17
Minus trades: 10
Zero trades: 0
Gross +R: +22.8183R
Gross -R: -9.4532R
Net R: +13.3651R
Average R per trade: +0.4950R
Winrate: 62.96%
TP: 13
SL: 8
MAX_HOLD: 6
Blocked trades only:
Trades: 27
Plus trades: 7
Minus trades: 19
Zero trades: 1
Gross +R: +5.8785R
Gross -R: -19.5849R
Net R: -13.7064R
Average R per trade: -0.5076R
Winrate: 25.93%
TP: 3
SL: 18
MAX_HOLD: 6
The filter improved the Week 27 result by +13.7064R.
Base Net R: -0.3412R
Filtered Net R: +13.3651R
Improvement: +13.7064R
Blocked Net R: -13.7064R
This is a strong historical filter result.
Why This Finding Matters
The M15 directional half filter is important because it did not simply remove all activity.
It kept 27 trades out of 54.
The filtered group preserved most of the gross positive result:
Base Gross +R: +28.6968R
Filtered Gross +R: +22.8183R
At the same time, it removed a large part of the gross negative result:
Base Gross -R: -29.0381R
Filtered Gross -R: -9.4532R
This is exactly the type of filter the research process is looking for.
A useful filter should reduce negative exposure while keeping enough positive potential.
The M15 directional half filter passed this historical Week 27 test.
Additional Supporting Signals
The audit also showed several related M5 conditions that supported the same general idea.
For example, when the M5 body was not against the candidate direction, the result was also positive.
m5_dir_body_against = NO
Trades: 26
Plus trades: 16
Minus trades: 10
Gross +R: +21.1843R
Gross -R: -9.3120R
Net R: +11.8723R
Another positive group was:
m5_dir_body_positive = YES
Trades: 24
Plus trades: 15
Minus trades: 9
Gross +R: +20.0761R
Gross -R: -8.2745R
Net R: +11.8016R
These results show that entry candle quality was important for EURJPY during Week 27.
However, EURJPY V1.28 will not add multiple filters at once.
The next version will add only the strongest and simplest filter from this audit:
M15 directional half filter.
Why M15 Directional Half Was Chosen First
The M15 directional half filter was chosen because it produced the strongest clean improvement in the historical Week 27 test.
The filtered result was:
Net R: +13.3651R
This was stronger than the M5 body filter candidate:
m5_dir_body_against = NO
Net R: +11.8723R
Both filters were useful, but the M15 directional half filter was the better first controlled change for EURJPY V1.28.
The goal of V1.28 is not to optimize everything at once. The goal is to test one clear filter in live paper observation.
Why Only One Filter Was Added
Only one filter is added to EURJPY V1.28 because the next test must remain controlled.
Adding several filters at the same time would make it difficult to understand which filter improved or damaged the result.
The selected filter is simple:
For BUY candidates:
allow only if the M15 close is in the upper half of the M15 candle.
For SELL candidates:
allow only if the M15 close is in the lower half of the M15 candle.
If this condition is not met, the entry is blocked.
This directly targets the strongest negative group found in the Week 27 audit.
Decision After Week 27
EURJPY V1.27 did not pass the +5R weekly target in raw form.
However, it produced enough gross positive R to justify a next controlled test.
The main decision after Week 27 is:
Create EURJPY V1.28.
EURJPY V1.28 will keep the V1.27 base logic and add one M15 directional half filter.
EURJPY V1.28 Description
New version: EURJPY V1.28 M15 Directional Half Filter
Base version: EURJPY V1.27 Soft Parallel 30/70
Symbol: EURJPY
Timeframe: M5
Mode: PAPER_SIM_ONLY
Orders sent: NO
Main change:
Block entry if the M15 close is not in the directional half for the candidate trade direction.
What Remains the Same in V1.28
EURJPY V1.28 keeps the V1.27 branch logic.
The following elements remain unchanged:
EURJPY symbol
M5 signal timeframe
Soft Parallel 30/70 logic
Single-position structure
Allowed parallel corridor types
Paper-only mode
No real orders
orders_sent=NO
Weekly target: +5R net
What Changes in V1.28
Only one new filter is added.
New filter:
M15 Directional Half Filter
Rule:
If candidate direction is BUY, allow only if the latest M15 close is in the upper half of the M15 candle.
If candidate direction is SELL, allow only if the latest M15 close is in the lower half of the M15 candle.
Blocked reason:
M15_CLOSE_NOT_IN_DIRECTIONAL_HALF
Status message:
SIGNAL_BLOCKED_M15_DIRECTIONAL_HALF
This filter is designed to block the specific losing group found during the EURJPY V1.27 Week 27 audit.
Week 28 Research Task
The Week 28 task is to observe EURJPY V1.28 as a separate EURJPY branch.
The test should answer the following questions:
Does the M15 directional half filter reduce Gross -R in live paper observation?
Does EURJPY V1.28 preserve enough Gross +R?
Does EURJPY V1.28 reach +5R net per week?
Does the number of SL exits decrease?
Does the trade count remain acceptable?
How many signals are blocked by M15_CLOSE_NOT_IN_DIRECTIONAL_HALF?
Are blocked signals mostly bad trades, neutral trades, or missed winners?
The most important Week 28 metrics are:
Gross +R
Gross -R
Net R
TP count
SL count
MAX_HOLD count
Blocked M15 directional half signals
Trade count after filter
Success Criteria for V1.28
EURJPY V1.28 will be considered successful if it shows the following:
Net R reaches or exceeds +5R for the week.
Gross -R is materially lower than in V1.27.
Gross +R remains strong enough after filtering.
SL count decreases.
The robot still produces enough trades for weekly evaluation.
Blocked signals show a logical reason and are not mostly missed winners.
orders_sent remains NO.
The ideal result is not necessarily maximum trade count.
The ideal result is a cleaner EURJPY branch with lower negative exposure and enough positive R to reach the weekly target.
Failure Criteria for V1.28
EURJPY V1.28 will be considered weak if one or more of the following happens:
Net R remains below 0R.
Net R remains far below +5R.
The filter removes too much Gross +R.
Gross -R remains high.
SL count remains high.
Trade count becomes too low for useful evaluation.
Blocked signals include too many trades that would have been strong winners.
The balance path remains unstable.
Any safety rule is violated.
If V1.28 fails, the EURJPY branch will need another audit before the next version is created.
Research Note
EURJPY V1 is treated as an independent research branch.
The V1.28 filter is added because the EURJPY V1.27 audit itself found a clear losing condition.
The filter is not copied from EURGBP. It is selected from EURJPY’s own Week 27 trade set.
This is important because different symbols can require different filters. EURGBP showed strong evidence for an M5 body filter. EURJPY showed stronger evidence for an M15 directional half filter.
EURJPY V1.28 must therefore be evaluated on its own Week 28 performance.
Safety Rules
EURJPY V1.28 remains paper-only.
Safety settings:
Mode: PAPER_SIM_ONLY
Orders sent: NO
No order_send execution
No live trading
No real account execution
No demo-stage promotion without separate review
All Week 28 results must be interpreted as research data only.
Final Decision
EURJPY V1.27 is not accepted as a demo-stage candidate after Week 27.
The raw version finished close to breakeven and below the weekly +5R target.
Final raw result:
Net R: -0.3412R
However, the Week 27 audit found a strong and simple filter candidate.
The historical M15 directional half filter test improved the same trade set to:
Net R: +13.3651R
Therefore, the EURJPY branch continues with a controlled next version:
EURJPY V1.28 M15 Directional Half Filter
The purpose of V1.28 is to test whether the M15 directional half filter can reduce negative exposure in live paper observation while keeping enough positive R to reach the weekly target.
Next Report
The next report should review EURJPY V1.28 after Week 28 live paper observation.
The report should include:
Total trades
Gross +R
Gross -R
Net R
TP count
SL count
MAX_HOLD count
Blocked M15 directional half signals
Comparison between allowed and blocked signal groups
Decision whether V1.28 continues, changes, pauses, or becomes a stronger robot candidate